<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Colin Reed on Medium]]></title>
        <description><![CDATA[Stories by Colin Reed on Medium]]></description>
        <link>https://medium.com/@creed4677?source=rss-7c07f4a4f0a5------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*Jnt8dcXrcoL4K7K3GVEv6Q.jpeg</url>
            <title>Stories by Colin Reed on Medium</title>
            <link>https://medium.com/@creed4677?source=rss-7c07f4a4f0a5------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 20:23:25 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@creed4677/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[The Number on the Screen and the Number on the Floor]]></title>
            <link>https://medium.com/supply-chains/the-number-on-the-screen-and-the-number-on-the-floor-ee4415dfa75b?source=rss-7c07f4a4f0a5------2</link>
            <guid isPermaLink="false">https://medium.com/p/ee4415dfa75b</guid>
            <category><![CDATA[supply-chain]]></category>
            <category><![CDATA[operations-management]]></category>
            <category><![CDATA[labor]]></category>
            <category><![CDATA[logistics]]></category>
            <category><![CDATA[warehouse-management]]></category>
            <dc:creator><![CDATA[Colin Reed]]></dc:creator>
            <pubDate>Wed, 18 Nov 2026 19:15:37 GMT</pubDate>
            <atom:updated>2026-11-19T15:42:51.347Z</atom:updated>
            <content:encoded><![CDATA[<h4>What your labor management system may or may not see, from an operator who ran one in a union distribution center.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*CVuaBsXDKxpbmS4Z.png" /></figure><p>It’s Super Bowl Sunday, and the night shift starts in a few hours. Fifty-five selectors are on the schedule. Eighteen have called in sick, most with less than three hours’ notice.</p><p>That’s a third of the selection workforce gone before the first pallet is built.</p><p>It wasn’t a surprise. In the union food distribution building I ran in San Diego, sick calls spiked on the holidays the contract didn’t pay: Thanksgiving weekend and the Super Bowl, the biggest of them all. You could predict the night. You couldn’t staff your way out of it.</p><p>The labor management system saw it too. Its workforce orchestration plan, the module that moves selectors across the freezer, cooler, and dry zones based on live productivity, told me selection would finish at 10:30 in the morning.</p><p>Selection needs to be done between 3 and 5 a.m. The first trucks leave the dock at 4:30. Finish at 10:30 and you get drivers waiting on selectors, a jammed dock, and late routes all over the county.</p><p>So we did what the plan couldn’t do on its own. We called two sister divisions, each about 75 miles away. They were Teamsters too, different locals, same union. We paid a premium to bring selectors in for the shift. We paid them time and a half, plus the 150-mile round trip, to save our own crew’s overtime and get the trucks out.</p><p>And for the rest of that night, the numbers on the screen drifted away from the numbers on the floor.</p><p>In theory, the system could have caught up. Borrowed selectors, contract labor, and schedule changes can all be keyed in as exception data, and a disciplined team can make the LMS see nearly everything. In practice, at 1 a.m. with a third of your crew missing and trucks due out in a few hours, nobody is sitting at a terminal entering exceptions. They’re on the floor.</p><h3>What the industry promises</h3><p>The pitch for labor management systems has been steady for twenty years. <a href="https://dcvelocity.com/articles/22361-what-have-you-done-for-me-lately">AMR Research put the typical gain at 10 to 30 percent</a>. Charles Zosel of Tom Zosel Associates, a labor management consultancy, put it at <a href="https://www.mmh.com/article/labor_management_software_improves_productivity/news">20 to 30 percent more effective than a site without one</a>.</p><p>The case studies come in lower. Manhattan Associates’ <a href="https://www.casestudies.com/company/manhattan-associates/case-study/giant-eagle-reduces-costs-and-increases-customer-value-with-manhattan-solutions">Giant Eagle deployment</a> reported 8 to 10 percent in its health and beauty DC and 3 to 5 percent in the others. Blue Yonder’s <a href="https://edge.sitecorecloud.io/blueyonderie8e6-corporateefb0-prod6ccb-be06/media/project/blueyonder/corporate/files/customers/americold.pdf">Americold case</a> reported 5 to 10 percent lower labor cost in certain locations.</p><p>Those are real gains, and I’m not dismissing them. A building with no standards that adds good ones should see a jump. But the vendors’ own people say where it comes from. A RedPrairie labor management executive put it plainly: “the software alone is not the real driver.”</p><p>I agree. In my experience, three other things set the number: the contract, the crew, and the incentive.</p><h3>The standard lives in the contract</h3><p>In a union distribution center, the engineered standard isn’t a software setting. It’s written into the collective bargaining agreement, usually for three to five years. <a href="https://www.mwpvl.com/html/lms.html">MWPVL International notes</a> that in union DCs, standards have to be negotiated with and accepted by the union.</p><p>Our selectors were held to 85 percent of standard for the day and 95 percent for the week. A selector who finished a shift under standard had to make up the shortfall on another shift, so the week still came in at 95.</p><p>So you don’t re-engineer the building whenever the data drifts. Either side can open it. The union can request a re-engineer, or the company can start one. In practice, it takes a trigger: a big swing in volume, an infrastructure change, a move to a new building, or an innovation that changes how selection and throughput are computed.</p><p>Can you raise the standard in the middle of a ratified contract? Technically, yes, through a memorandum of understanding (MOU), a side agreement both parties sign. In practice, almost no union agrees to a higher standard mid-contract without getting something back: incentive pay, changes to sick days, or softer discipline rules. The concession usually eats most of the gain the re-engineer was supposed to deliver.</p><p>Distribution is well suited to this. Case dimensions change, but the work doesn’t. You’re still driving a pallet jack, pulling cases from a slot, and building a pallet. Airline catering is the opposite. At Gate Gourmet, menu cycles and airline specs changed so often from building to building that a standard couldn’t keep up. Software vendor <a href="https://softeon.com/?p=3073">Softeon has observed the same pattern</a>: interest in full engineered standards holds up in grocery and foodservice and fades elsewhere.</p><h3>The case: same building, same slots, 20 percent more</h3><p>My building had no productivity incentive. The only incentive in the contract was not to fall below standard.</p><p>During a labor disruption, we staffed the floor with crews from other buildings. Some of them worked under selection incentives at home. All of them were on strike-coverage pay, which runs well past time and a half once you add hotels, meals, and per diem. Strike coverage is expensive. Their incentive was simple: finish the work and go home.</p><p>We expected throughput to drop. For a day or two it did, while they learned the pick path, the temperature zones, and how we wanted pallets placed. Then the night crews outran our regular workforce. Throughput rose 20 percent.</p><p>Same building. Same slots. Same software. A 20 percent swing, about as much as the entire promised LMS gain, and it came from the crew and the incentive.</p><p>The following week, executive leadership called me directly to ask if I was really reporting those numbers. They suspected a clerical error, or someone fudging the data. I told them it was real. They accepted that, and then asked the question that matters: how do we make this real all the time, labor stoppage or not?</p><p>That’s the rub. Without the right alignment of incentives, you can’t.</p><h3>The data cuts both ways</h3><p>The common fear about LMS in a union building is that it’s a weapon. It isn’t, for either side.</p><p>We used it for coaching and for discipline. For new hires, it was a progress report, not a hammer: here’s where you are, here’s where you need to be. We posted the previous night’s numbers on the briefing board every day and the week’s numbers every Sunday night. Selectors checked them, because a shift under standard had to be made up with a stronger shift before the week closed.</p><p>Stewards used the same data. Before filing a grievance over discipline, they’d ask us to pull the LMS record to see whether the employee was telling the truth.</p><p>That’s dual accountability. Management doesn’t always enjoy having its own data turned on it, but it keeps supervisors honest. Because both sides can pull the lever, neither side can own it.</p><h3>The fix: hybrid decisions, and a price for the standard</h3><p>On the Super Bowl night, the LMS forecast was built for a building that no longer existed: our crew, our pick paths, our rates. Borrowed selectors rarely get fed cleanly into the system in the middle of a crisis. So for a stretch, the system is working from incomplete data.</p><p>That doesn’t mean you make bad decisions. It means the decision becomes a hybrid of what the machine says and what experienced supervisors see on the floor. Some events, by definition, can’t be modeled. The job is still to dispatch on time, safely, while onboarding people who know the company’s processes but not your floor.</p><p>The other half of the fix is treating the standard itself as an asset with a maintenance cost. Here’s roughly what that costs:</p><ul><li><strong>A one-time build by an outside firm.</strong> For a 300,000-square-foot DC with a dozen direct processes, Easy Metrics, which sells a cheaper data-driven alternative, <a href="https://www.easymetrics.com/wp-content/uploads/2023/04/Data-Driven-Labor-Standards-v9.pdf">puts engineering at upwards of $75,000 to $100,000</a> and several months. Consultancy Jackpine’s example of a full standards-and-LMS program for a 500,000-square-foot building <a href="https://jackpine.us/2025/10/13/understanding-the-real-roi-of-engineered-labor-standards/">runs $300,000</a>.</li><li><strong>A re-engineer on the contract cadence.</strong> Firms like West Monroe, enVista, and 4SIGHT do this work. A re-engineer is narrower than a first build, but you buy it again every contract, and in a union building it comes with a negotiation. Even if it cost as much as the first build, spread over a four-year contract that’s about $19,000 to $25,000 a year.</li><li><strong>More often than the contract allows.</strong> Seasonal re-slots, new SKUs, and mix shifts don’t wait for bargaining. If you’re re-studying processes several times a year, you’re paying consultants repeatedly for the same building.</li><li><strong>Insourced.</strong> Posted salaries for distribution center industrial engineers run about $70,000 to $120,000, and senior labor-management engineers about $130,000 to $140,000, before benefits. C&amp;S Wholesale Grocers runs its standards program with an <a href="https://www.career.com/job/c-s-wholesale-grocers/senior-industrial-engineer/j202606050950230529462">in-house engineering team</a>.</li></ul><p>For one building, the contract cadence and an outside firm make sense. For a network, an in-house engineer pays for itself.</p><h3>What it means for the people writing the checks</h3><p>If you’re looking at a distributor, you don’t need to read 55 collective bargaining agreements. You need a few answers in aggregate.</p><p>First, how much of the network is union? Fifty-fifty, sixty-forty, seventy-thirty?</p><p>Second, within the union share, how many locations will struggle to deliver the gains in the plan? And of those, which are habitually poor performers, and which are boxed in structurally by what the contract allows?</p><p>Those are different problems. A poor performer can be fixed with management. A structural limit can’t be fixed until the contract opens. So the last question is the one that sets the timeline: when does each of those contracts come up for renegotiation, when the structural limits can actually be addressed?</p><p>The same logic applies to the newest pitch, standards that machine learning recalibrates continuously. In a union building, real-time recalibration has to be negotiated ahead of time, at the contract table. Bolting it onto an existing agreement would take a concession to offset it, and the concession can wipe out the gain of doing it at all.</p><h3>What would change my mind</h3><p>The union, and the Teamsters in particular, coming to see itself as an extension of the company’s earnings rather than an opponent across the table on wages.</p><p>If labor and management shared an understanding of what innovation could produce, including kinds of work that don’t exist in the operation yet, the incentives could line up: the company, the shareholders, the management team nationally and locally, and the union across the network and in each contract. Greener pastures may well lie on the other side of the fence. Someone just has to take the fence down.</p><p>That isn’t the reality today. What I saw instead was resistance, with every technology change read as a move against the people on the floor. Until that changes, the contract will set the number, whatever the screen says.</p><p><em>Colin Reed is a former food distribution and airline catering operator. He writes about operations at </em><a href="https://modernwealthmodel.com/ai-operations/"><em>Modern Wealth Model</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ee4415dfa75b" width="1" height="1" alt=""><hr><p><a href="https://medium.com/supply-chains/the-number-on-the-screen-and-the-number-on-the-floor-ee4415dfa75b">The Number on the Screen and the Number on the Floor</a> was originally published in <a href="https://medium.com/supply-chains">Supply Chains</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[It Feels Like Cheating. It Isn’t.]]></title>
            <link>https://medium.com/@creed4677/it-feels-like-cheating-it-isnt-f42cebfc948a?source=rss-7c07f4a4f0a5------2</link>
            <guid isPermaLink="false">https://medium.com/p/f42cebfc948a</guid>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[ethics]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[future-of-work]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Colin Reed]]></dc:creator>
            <pubDate>Tue, 17 Nov 2026 18:49:03 GMT</pubDate>
            <atom:updated>2026-11-17T18:49:03.058Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="A dim living room at night seen from the couch, with a sports match on a wall-mounted TV and warm lamps on either side" src="https://cdn-images-1.medium.com/max/1024/0*DGXr5S9EDdXv_I6b" /><figcaption>Photo by <a href="https://unsplash.com/@momina76?utm_source=medium&amp;utm_medium=referral">Momina Zeshan</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p><em>The advantage of being early to AI agents feels unethical. Here’s why it’s not, and why the window won’t stay open.</em></p><p>At 8:41 on a Thursday night in November 2026, I was on the couch watching the Avalanche play the Predators. My cat was asleep on my leg. I was talking into my phone.</p><p>While I talked, one of my AI agents took a half-formed idea, wrote it up in full, mapped it to every outlet I publish on, and handed it to a second agent that runs my websites. Then it went straight back to the research project it had been working on before I interrupted. Same agent, same few minutes, three jobs in parallel, while I watched a hockey game.</p><p>I’ve had a nagging feeling about that kind of moment for a while, and I want to name it honestly. It feels like I’m getting away with something.</p><h3>The feeling</h3><p>People who already understand what agents can do have an advantage over people who don’t that is so lopsided it almost feels illegal. If not illegal, then unethical. It feels like walking an ethical fence just because you showed up early.</p><p>I take that feeling seriously, because a lot of people share it from the other side. If you aren’t using these tools yet, watching someone run five workstreams from a couch looks less like skill and more like a loophole.</p><h3>What’s actually happening</h3><p>Here’s the thing, though. Nothing the agents do is new work. Every task is something a normal person could do under normal conditions. Research a topic. Draft a page. Check a fact against the source. Schedule a post. Update a directory. People have always done all of that.</p><p>What’s different is the form, not the substance. The same work now runs a hundred times faster, concurrently, across several workstreams at once, around the clock. There are no breaks, no meals, no sleep, and no pushback.</p><p>That isn’t espionage, and it isn’t cheating. It’s a new form factor.</p><h3>We’ve done this before</h3><p>I learned cursive in sixth grade. In seventh and eighth grade, I went to computer lab and learned home row on those giant Macs that looked like colorful bubbles and took up three feet of desk in every direction. Typing was the new form factor. Within a few years cursive was mostly gone.</p><p>The smartphone was the next one. Now we’re in another: you talk to the computer and it does the work for you. I barely type anymore. I’d bet that typing, as the main way most people work, looks a lot like cursive within five years.</p><p>At each of those shifts, the early people looked like they had an unfair edge. The kid who could type 70 words a minute in 1996 wasn’t cheating. He just got to the new form factor first.</p><h3>Why it feels unethical anyway</h3><p>I think the guilt comes from two things.</p><p>The first is the size of the gap. When one person can do in an evening what used to take a team a week, the difference stops feeling like skill and starts feeling like privilege. But the tools are available to almost anyone with a phone and a subscription. The gap is in attention and willingness, not access.</p><p>The second is that the advantage is temporary, and we can feel that. Early adopters capture outsized returns precisely because most people haven’t moved yet. That’s how every new tool has worked. The returns compress as everyone catches up. That isn’t a moral failing. It’s how prices tell people where effort is worth spending.</p><h3>The part that isn’t comfortable</h3><p>There is a hard edge here, and I don’t want to skip it. People who don’t learn to direct agents are going to be displaced and disrupted in a very big way. That’s my honest read. You don’t have to like it for it to be true, and pretending otherwise doesn’t help anyone.</p><p>The useful response isn’t guilt. It’s to say out loud what the work is and to show people how to do it. That’s a big part of why I write this stuff in public, with the receipts.</p><h3>The life it makes possible</h3><p>Back to the couch. What struck me that night wasn’t the productivity. It was that the productivity didn’t cost me the evening.</p><p>You can live a comfortable life and work at the same time. You can bring work fully into your day-to-day, as long as you stay engaged and keep talking to your agents. That’s not a promise of leisure. It’s a different shape for the workday, one where directing the work replaces doing every step of it.</p><p>It’s a wild time to be alive. The ethical question isn’t whether to use the advantage. It’s what you do with the time it gives back.</p><p>Sound Operations · Owner’s Desk. Get <a href="https://modernwealthmodel.com/?utm_source=medium&amp;utm_medium=referral&amp;utm_campaign=sound-operations">Sound Ops</a>, free every other Wednesday, at modernwealthmodel.com.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f42cebfc948a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Laziest Retirement Contribution I Make]]></title>
            <link>https://medium.com/the-alternative-ledger/the-laziest-retirement-contribution-i-make-013465d98fe8?source=rss-7c07f4a4f0a5------2</link>
            <guid isPermaLink="false">https://medium.com/p/013465d98fe8</guid>
            <category><![CDATA[side-hustle]]></category>
            <category><![CDATA[retirement]]></category>
            <category><![CDATA[passive-income]]></category>
            <category><![CDATA[bitcoin]]></category>
            <category><![CDATA[personal-finance]]></category>
            <dc:creator><![CDATA[Colin Reed]]></dc:creator>
            <pubDate>Tue, 17 Nov 2026 15:24:39 GMT</pubDate>
            <atom:updated>2026-11-17T15:24:39.316Z</atom:updated>
            <content:encoded><![CDATA[<h4>A background app sells my spare bandwidth. The payout goes to cold storage. Here’s what that adds up to, and what I’d tell you before you install it.</h4><figure><img alt="Grid card: $60 a month into Bitcoin, compounded monthly, at 0, 10, 20, 30, and 40 percent a year over 2, 3, 5, 10, 15, and 30 years. At 10 percent, 30 years of $21,600 in contributions grows to about $124,758. The 30-year values at 30 and 40 percent are marked as a thought experiment, not a forecast." src="https://cdn-images-1.medium.com/max/1024/1*dOtKUh6ECI6PlO-yodOynQ.png" /></figure><p>There’s an app running on my computer right now that I haven’t thought about in days. It’s Honeygain. It sells a slice of my home internet bandwidth, the capacity I’m paying for and not using, to businesses that need to see the web from a residential connection in my area. I get paid for the traffic that passes through.</p><p>It isn’t glamorous. Every so often the connection drops and I have to restart the app. Otherwise it runs in the background and never asks for anything.</p><p>It pays me somewhere between $50 and $75 a month. Thirteen days into November 2026, I’m already at $30. That isn’t life-changing money. But I stopped thinking of it as spending money a while ago and started thinking of it as a contribution. Every payout goes to a Bitcoin cold storage wallet, and I never touch it.</p><p>That shift, from “small side income” to “automatic monthly contribution,” is the whole point of this piece.</p><h3>What Honeygain actually does</h3><p>Honeygain describes its business as routing traffic for vetted partners doing things like price comparison, ad verification, SEO, and public web data collection. When a partner needs to load a page as if from your region, the request goes out through your connection. Honeygain says only the connection is shared, not your personal data, and it pays by the volume of traffic that passes through.</p><p>Two honest notes before anyone installs it. First, you are letting other companies’ traffic exit through your home IP address. Read the terms, and check your internet provider’s terms too. Second, Honeygain’s own estimate for one device in the US is roughly $3 to $10 a month. My earnings run well above that, so treat my number as one person’s result, not a promise. Yours will depend on where you live, how many devices you run, and how much demand there is in your area.</p><h3>Where JumpTask fits</h3><p>JumpTask is the payout side. It’s an earning platform with small paid tasks, and since August 2026 it’s also where Honeygain sends gift card and crypto payouts. My Honeygain earnings feed straight into JumpTask, where they’re held as credits at a fixed rate of 1,000 credits to the dollar. From there you cash out through a payout partner called GIFQ for gift cards, prepaid cards, PayPal, or crypto, including Bitcoin sent to an address you choose. That can be a hardware wallet, which is what I use.</p><p>JumpTask used to run on its own token, JMPT. It dropped JMPT support in August 2026 and moved everyone to the credit system. If you read older reviews that call it a crypto or DeFi platform, that’s out of date.</p><p>The minimum is low: JumpTask lets you cash out from 50 cents. The part to plan around is time. JumpTask told me to expect approval in 24 to 72 hours. My last cash-out took 12 days. That’s fine for money you’re not planning to spend, which is exactly the point.</p><h3>The grid</h3><p>Here’s the question that made this interesting to me. What happens if $60 a month from an app I forget about goes into Bitcoin every month and stays there?</p><p>I can’t tell you what Bitcoin will return. Nobody can. So the grid below shows a range of compound annual growth rates, from 0 percent (you only keep what you put in) to 40 percent. Each cell assumes $60 goes in at the start of every month and compounds monthly at that annual rate.</p><pre>Years  Put in        0%        10%        20%         30%          40%<br>2      $1,440    $1,440     $1,593     $1,751      $1,914       $2,083<br>3      $2,160    $2,160     $2,510     $2,897      $3,321       $3,784<br>5      $3,600    $3,600     $4,630     $5,922      $7,527       $9,501<br>10     $7,200    $7,200    $12,087    $20,659     $35,473      $60,598<br>15    $10,800   $10,800    $24,097    $57,328    $139,235     $335,411<br>30    $21,600   $21,600   $124,758   $940,575  $7,266,104  $52,514,693</pre><p>Read the bottom-right corner with a straight face, and then don’t believe it. Thirty years at 30 or 40 percent a year is a thought experiment, not a forecast. Growth rates that high don’t hold for decades in any asset, and a Bitcoin that grew that much would be a very different thing in a very different world. The rows I’d actually take seriously are the near ones, and the left half of the table.</p><p>What the grid does show is how much time does the work. At 10 percent, the 30-year row turns $21,600 of contributions into about $125,000. Most of that came from leaving it alone, not from the app.</p><h3>Why I think about it this way</h3><p>Most people treat side income as found money, and found money gets spent. The trick here isn’t Honeygain. It’s routing a small, automatic stream somewhere you won’t touch it, in an asset you’ve chosen to hold for a long time, and then forgetting about it.</p><p>Bandwidth is just the stream I happened to have. It costs me nothing extra, it takes almost no attention, and it pays into the same place every month. That’s the kind of money that’s easy to undervalue, because it never shows up as a decision.</p><h3>The limits</h3><p>The income is small and not guaranteed. It depends on demand in your area, it drops when the app disconnects, and the payout platform has changed its rules before and will probably change them again. Cash-outs can take days or weeks to clear. You’re sharing your home connection with businesses you’ll never see. And the grid is arithmetic, not a prediction: Bitcoin can fall hard and stay down for years, and the 0 percent column is a real possibility.</p><p>None of that changes the basic idea. A small automatic contribution you never touch is worth more than it looks. Just go in knowing which part is the app and which part is time.</p><p><em>This isn’t financial advice. Nobody paid for this piece, and I have no affiliation with Honeygain, JumpTask, or GIFQ.</em></p><p>Sound Operations · Money. Get Sound Ops, free every other Wednesday, at <a href="https://modernwealthmodel.com/?utm_source=medium&amp;utm_medium=referral&amp;utm_campaign=sound-operations">modernwealthmodel.com</a>.</p><h3>Make Better Financial Decisions With Your Own Numbers</h3><p>At <strong>Investor’s Handbook</strong>, we turn real investing and money questions into practical frameworks and <strong>free downloadable Excel models</strong> you can use to evaluate your own decisions.</p><p>Join our free newsletter for the Tuesday Digest, Thursday Weekly Framework, and a growing library of financial tools covering investing, retirement, debt, and building wealth.</p><p>📩 <a href="https://investorshandbook-newsletter-c72f94.beehiiv.com/"><strong>Get Free Financial Models &amp; Join Investor’s Handbook</strong></a></p><p><em>Learn the framework. Run the model. Make the decision.</em></p><p>🎥 <a href="https://www.youtube.com/channel/UCLcEzBcpLsD8cDUZGG246DQ"><strong>Watch on YouTube</strong></a> · 𝕏 <a href="https://x.com/InvestorsHndbk"><strong>Follow on X</strong></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=013465d98fe8" width="1" height="1" alt=""><hr><p><a href="https://medium.com/the-alternative-ledger/the-laziest-retirement-contribution-i-make-013465d98fe8">The Laziest Retirement Contribution I Make</a> was originally published in <a href="https://medium.com/the-alternative-ledger">The Alternative Ledger</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Questions Show Up in the News Before They Show Up on the Calls]]></title>
            <link>https://medium.com/@creed4677/the-questions-show-up-in-the-news-before-they-show-up-on-the-calls-c62a8bc8baaf?source=rss-7c07f4a4f0a5------2</link>
            <guid isPermaLink="false">https://medium.com/p/c62a8bc8baaf</guid>
            <category><![CDATA[logistics]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[economics]]></category>
            <category><![CDATA[supply-chain]]></category>
            <category><![CDATA[food-industry]]></category>
            <dc:creator><![CDATA[Colin Reed]]></dc:creator>
            <pubDate>Thu, 05 Nov 2026 17:28:46 GMT</pubDate>
            <atom:updated>2026-11-10T18:47:42.650Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="Modern Wealth Model Agentic Operations graphic reading The Questions Show Up in the News First, about foodservice distribution earnings week and AI agents" src="https://cdn-images-1.medium.com/max/1024/0*jd4z8i9MgoPLsaIQ.png" /></figure><p>In <a href="https://modernwealthmodel.com/expert-network-compliance-ai-tools/">Part 1</a>, I wrote about a weekly digest I built that reads public sources and turns the prior week into a five-minute briefing. In <a href="https://modernwealthmodel.com/expert-knowledge-capture-ai-briefing/">Part 2</a>, I wrote about the folder of what I know that sits behind it. This part is about what the digest has been reading, because the pattern is starting to say something about the questions people are about to ask.</p><p>The digest sorts every item into eight lanes: broadline foodservice distribution, AI software in distribution, AI infrastructure, airline catering, last-mile, freight, robotics and automation, and restaurants as demand signals. Each item also gets a few theme tags. After three weekly issues, from October 12 to November 1, 2026, the archive holds 54 items.</p><h3>What three weeks of tags say</h3><p>The top four tags are:</p><ul><li>Earnings: 16 items</li><li>AI agents: 8</li><li>Consumer demand: 8</li><li>AI planning, routing, and pricing: 6</li></ul><p>Thirteen of the 54 items touch AI directly, through agents, planning, warehouse systems, or fleet telematics. The names that came back in more than one week were Sysco, US Foods, Samsara, DoorDash, Instacart, AWS, and the airlines that buy catering.</p><p>Three weeks isn’t a trend. It’s a starting read, and I’ll treat it that way. But two of the top four themes are AI, and the third, consumer demand, is the thing AI forecasting is supposed to see first. That lines up with what I hear asked.</p><h3>Earnings week: what the broadliners said</h3><p>This piece is going up the week Sysco, Performance Food Group, and US Foods all report. Every figure below is the company’s own, from its release.</p><p><a href="https://investors.sysco.com/annual-reports-and-sec-filings/news-releases/2026/11-03-2026-130507019">Sysco</a> reported on November 3 that local case volume in its U.S. Foodservice segment rose 2.7% and total U.S. Foodservice cases rose 2.6%, while that segment’s gross margin widened 14 basis points to 19.2%. Its AI number is still a target, not a result: the company says $100 million of AI-enabled cost savings is built into its fiscal 2027 guidance, with the benefit weighted to the second half, and it is still preparing to close the Restaurant Depot deal.</p><p><a href="https://investors.pfgc.com/press-releases/press-release-details/2026/Performance-Food-Group-Company-Reports-First-Quarter-Fiscal-2027-Results/default.aspx">Performance Food Group</a> reported on November 4 that independent Foodservice cases rose 7.8%, or 5.5% excluding acquisitions, against total case growth of 4.2%. Gross profit grew 8.5% on 5.5% sales growth, which the company tied in part to selling more independent and house-brand cases. The release made no AI or routing claim.</p><p><a href="https://ir.usfoods.com/financials/quarterly-results/default.aspx">US Foods</a> reported on November 5 that independent restaurant cases rose 5.0% while chain volume fell 1.5%, for total case growth of 1.3%. Adjusted gross profit rose 5.7% on those 1.3% more cases, which works out to roughly 4% more gross profit per case. The company credited its digital tools and AI initiatives with sales growth, productivity gains, and margin expansion, a claim the release did not put a number on.</p><p>What I read first in a print like this isn’t revenue. In an inflationary quarter, dollar growth flatters everyone. I look at cases, gross profit dollars per case, and the mix between independent restaurants and chains. Those tell you whether a distributor is winning business or just passing through higher prices.</p><h3>AI agents moved from pilot to paid</h3><p>The biggest shift in the AI lanes over these three weeks was vocabulary. Vendors stopped selling assistants and started selling agents that take approved actions.</p><p>In fleet telematics, Samsara, Netradyne, and Motive all shipped agents within about a week of each other in October. They can call drivers, open maintenance work orders, flag hours-of-service risk early, and take a truck out of service under rules the fleet sets. In warehouse and enterprise software, Manhattan Associates told investors on October 27 that its AI pilots were converting to paid subscriptions, and Oracle launched agents that plan load consolidation and allocate short supply, then act once a plan is approved.</p><p>None of them published pricing. The open question isn’t whether the agents work. It’s who signs off on what they’re allowed to do, who pays when one is wrong, and what they cost at renewal. In a unionized fleet, an agent calling a driver is <a href="https://modernwealthmodel.com/agentic-ai-warehouse-labor-constraint/">a labor-relations question</a> before it’s a software question.</p><h3>Demand is splitting, not falling</h3><p>Consumer demand was the other recurring theme, and the signals disagree in a useful way. Ecolab, which sells cleaning and sanitation programs to restaurants, said on October 27 that North American restaurant foot traffic was down by a mid-single-digit percentage. The same week, Chili’s reported comparable sales up 10.1% and Burger King U.S. reported 8.4%, while Wingstop fell 5.4%.</p><p>That isn’t a contradiction. A few chains winning on value and execution are taking traffic from everyone else. For a distributor, chain volume depends on which brands you carry, not on how the industry is doing on average. The independent restaurants, where broadliners earn their best margins, are where a traffic decline shows up first, in smaller orders.</p><p>Proteins split too. Chicken is getting cheaper on oversupply while beef keeps climbing, and those two move a distributor’s margin per case in opposite directions depending on how each contract is priced.</p><h3>AI routing: the honest numbers are small</h3><p>The AI results with real numbers behind them are modest. US Foods’ public claims, compiled in late October, include about a 2% gain in cases per mile from routing software and a 3% lift in e-commerce cart conversion. Its seller headcount rose at the same time, so the gain is more output per seller, not fewer sellers.</p><p>Two percent on a large fleet is real money. It’s also a long way from the transformation language around it. When a company attaches a savings target to AI, the question is which cost bucket moved, by how much, and what share of it the AI earned.</p><h3>Questions I’d expect next</h3><p>If the pattern holds, these are the questions that come next from investors, strategy firms, and operators:</p><ol><li>Is independent case growth real, or mostly inflation?</li><li>Which distribution workflows can an AI agent run without a human approving each action?</li><li>Will AI agents carry separate pricing at warehouse and ERP software renewals?</li><li>Is a 2% routing gain the ceiling or the floor?</li><li>Does one company owning both restaurant and grocery delivery (DoorDash agreed to buy Instacart on October 29) change what independents pay to reach customers?</li><li>Does a possible tariff on imported robotics and machinery change the timing of planned warehouse automation?</li></ol><h3>Where this read breaks</h3><p>Three weeks of tags is a small sample, and the tags are mine. I chose the lanes, I chose the sources, and the digest counts what it was built to notice. An earnings-heavy month will always look earnings-heavy. A quiet lane might be quiet because nothing happened, or because my source list missed it.</p><p>The bigger limit is that a count of headlines measures attention, not importance. AI agents got eight items in three weeks because vendors issued a lot of releases, not because eight distributors deployed them. The digest can tell me what is being talked about. It still takes someone who ran the dock to say which of it will matter.</p><p>Sound Operations · Operating Floor. Get <a href="https://modernwealthmodel.com/?utm_source=medium&amp;utm_medium=referral&amp;utm_campaign=sound-operations">Sound Ops</a>, free every other Wednesday, at modernwealthmodel.com.</p><p>Written by Colin Reed</p><p>Colin Reed is an operations executive and independent consultant whose career spans foodservice distribution, airline catering, last-mile delivery, and hospitality, from restaurants and bars to luxury resort food and beverage programs. He has held distribution P&amp;L with capital authority and run enterprise system selections across warehouse management, labor management, and ERP platforms, and advises funds, corporates, and management consultancies through expert networks and research firms. On Modern Wealth Model he writes on applied AI in the physical economy, Bitcoin, and long-horizon retirement planning — and builds the calculators.</p><p><em>Originally published at </em><a href="https://modernwealthmodel.com/foodservice-distribution-ai-agents-earnings-trends/"><em>https://modernwealthmodel.com</em></a><em> on November 5, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c62a8bc8baaf" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Six Builds, Retired. What the Graveyard Cost.]]></title>
            <link>https://medium.com/the-alternative-ledger/six-builds-retired-what-the-graveyard-cost-205e36a27e2e?source=rss-7c07f4a4f0a5------2</link>
            <guid isPermaLink="false">https://medium.com/p/205e36a27e2e</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[entrepreneurship]]></category>
            <category><![CDATA[side-hustle]]></category>
            <category><![CDATA[business]]></category>
            <category><![CDATA[money]]></category>
            <dc:creator><![CDATA[Colin Reed]]></dc:creator>
            <pubDate>Thu, 05 Nov 2026 15:05:37 GMT</pubDate>
            <atom:updated>2026-11-05T15:05:37.778Z</atom:updated>
            <content:encoded><![CDATA[<p>On March 9, 2026, my research aggregator ran seven times. Every run reported every site a success. Five of the six sites returned zero listings, every time.</p><p>The dashboard showed all of them green. Risk: low.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Uq4W0oxC592dGrY7.png" /><figcaption>The Research Aggregator&#39;s site health page, reopened from the archive in October 2026. Every site green, almost every one finding nothing.</figcaption></figure><p>That was the first of six things I built that spring and later shut down. I had time on my hands, AI tools that were getting remarkably good, and nothing stopping me from building whatever I could think of. The barrier to entry had dropped so far that I just started throwing spaghetti at the wall.</p><p>Here’s the wall.</p><h3>1. The research aggregator: faster than the job it replaced, until it wasn’t</h3><p>Before expert networks, I was working the research side: User Interviews, Respondent, L&amp;E Research, Recruit and Field, FocusGroups.org, and a dozen others. The pay is small, $15 to $35 a study, and the best ones reach $85 or $100. The real cost is the screener. Some run 10 questions, some run 30, and you answer every one for no guarantee.</p><p>So I built a machine to do the screening part. A scraper pulled listings from six sites on a schedule. A local dashboard sorted them by pay and time. A browser extension filled in screener answers from a saved profile. Fifty files in three days.</p><p>It looked great. It didn’t work. Every site had its own CAPTCHAs, bot monitoring, login walls, and tolerance for automation, and the questions varied too much for any stock set of answers. The tool kept stopping to ask me things. In the end it took more time than applying by hand, which was the one problem it existed to solve.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*hfnhPJ1NrQsNJ8gK.png" /><figcaption>The metrics page. Zero applications. The tool never got used for the thing it was built to do.</figcaption></figure><h3>2. The Vault: I stopped believing in the buyer before I finished the product</h3><p>The Vault was early Modern Wealth Model: five printable Bitcoin and retirement planning tools, a DCA tracker, a FIRE planner, and a few guides, listed on Gumroad at $4.99 each.</p><p>The AI tools of the time needed a lot of manual cleanup to get those files right. Thinking through the cleanup made me think about who would actually buy them. The longer I thought, the longer the listings sat unactivated. Eventually I admitted the idea was bad and closed it.</p><h3>3. The arb bot: the one everybody builds</h3><p>Every person who discovers agentic AI eventually builds a trading bot. It looks like free money that runs itself.</p><p>Mine started on Polymarket, which turned out to be off limits to US users. Trading there also meant a crypto wallet the bot would have to control, and that got complicated fast. So I rewrote it for Kalshi, which has a normal API key and runs from a Python script in Terminal. It ran on a 2018 MacBook Air.</p><p>The idea: some events have many outcomes and exactly one pays $1. If the cheapest “yes” on every outcome adds up to less than a dollar, buy one of each and keep the difference. Small, repeated, close to riskless. The bot checked the markets every five seconds.</p><p>Ten build sessions in, the log told the story. It attempted 113 of those trades. Every one came back with zero of its legs filled, out of anywhere from 23 to 96. Reading the code months later, the more likely explanation is that the orders were rejected outright: the bot could see the markets but probably never had permission to trade them.</p><p>Even working, it had a deeper flaw. Sending up to 96 separate orders one at a time means prices move between orders, and any leg that misses leaves you holding a plain bet. That’s the opposite of riskless.</p><p>I didn’t stop it because of one bad result. I stopped because it needed constant rework, someone going in to tell it to try something different. A bot that needs a babysitter isn’t passive income.</p><h3>4. The Nostr agent: shelved, not dead</h3><p>Nostr is a social network built on public and private keys, so your identity is yours and not the platform’s. Much of the early crowd came from Bitcoin, the self-custody and privacy people, folks like Matt Odell.</p><p>My plan was an agent that would post updates on what I was building and join relevant conversations, openly as a human-plus-AI account. I spent a stretch posting to myself on Nostr, setting the table for the agent to take over.</p><p>Then my attention went to things that paid off sooner, and the project drifted. The account and the posts are still there. I don’t think the network is growing fast enough yet to justify the time. Of the six, it’s the one I’d most like to revive, because agents, automation, and sound money look like they’re heading toward the same place.</p><h3>5. The dashboard for a business: one command too many</h3><p>This one grew out of my own Command Center, which started as an expense tracker and turned into the place where I run nearly everything: income, savings, retirement planning, projects, filings, and work metrics.</p><p>I figured businesses would want one, so I built a custom version for my sister as a pilot, with a setup guide. It ran locally and started from Terminal: open it, paste one line, press Enter.</p><p>That one line was too much. Most people have never opened Terminal, and the request alone stops them. I couldn’t sell anything that depended on it. The idea isn’t gone. It lives on as a workflow in my operations consulting practice instead of a product you install.</p><h3>6. WordPlay: Words with Benefits</h3><p>A friend and I have played Words With Friends for years, and for years we were both on the dating apps. One day we joked that the two should be one app: match with someone and, instead of a chat window, you get a Scrabble board. We called it Words with Benefits.</p><p>Once the tools could carry me, I built it. Four sessions got me a two-player board, tiles you could drag into place, a dictionary checking every word, and chat along the side. Next came profiles and photos, and next after that came a paid hosting plan for something I wasn’t sure anyone wanted.</p><p>So I paused it, my other projects took off, and WordPlay got left behind. The code still sits in a private repo.</p><h3>The bill</h3><p>All six together took roughly 32 to 62 hours by my count, about a week and a half of full-time work. That’s an estimate from file dates and session counts, not a time log.</p><p>What did that buy? Terminal. Python, JavaScript, and CSS. How to scrape, store, and display data on my own machine. How to sign API requests. How to tell a build that’s struggling from one that’s dead. Before AI tools, I couldn’t have learned any of that for a week and a half of anything. I couldn’t have learned it at all.</p><p>Not all of it was buried, either. The aggregator’s habit of rating platforms side by side became how I track my expert network work today. The Vault’s products came back as real offerings tied to my consulting practice, not third-party PDFs. And all of them taught me to build odd things from nothing, which I do every week.</p><h3>What didn’t get cheaper</h3><p>Almost everything got cheaper. Ideas got expensive.</p><p>When building costs close to nothing, the limit is how many things you can keep pointed in the right direction. As I write this, six of my projects have a to-do waiting on me. A seventh is waiting for my approval to act. Seven more are partway between steps, two pinned ones are stalled waiting on something, and nine more are pinned but quiet for the week. Underneath all of that, about 90 recurring jobs run on schedules.</p><p>I don’t have a kill rule. I sort by value now, in the medium term, or in the long run, and then against what I need to do, what I want to do, what’s just fun, and what might pay off in ways I can’t see yet. It’s fluid, and it has to be. The graveyard is what that sorting looked like before I had it.</p><h3>No project is half-built</h3><p>People talk about half-built projects. I don’t think they exist. A project is built, still building, or scrapped. “Half-built” just means you stopped, and either you stopped for a reason, which is a lesson, or you can pick it back up months later when you find the market, the model, or the missing piece.</p><p>Some ideas don’t have a home yet. Sometimes you have to watch and listen for a while before you realize what the thing was supposed to be.</p><p>One test for this week: list everything you started this year and mark the ones you’d pay to restart from scratch. My list is short. The Command Center, which I use constantly, is on it. None of the six above are, at least not yet.</p><p><em>Related: </em><a href="https://medium.com/the-alternative-ledger/i-said-20-hours-a-week-my-logs-say-6-to-16-ffa5214735d2"><em>I Said 20 Hours a Week. My Logs Say 6 to 16.</em></a></p><p><em>Sound Operations · Owner’s Desk. Get </em><a href="https://modernwealthmodel.com/?utm_source=medium&amp;utm_medium=referral&amp;utm_campaign=sound-operations"><em>Sound Ops</em></a><em>, free every other Wednesday, at modernwealthmodel.com.</em></p><p>📩 <a href="https://investorshandbook-newsletter-c72f94.beehiiv.com/"><strong>Join Investor’s Handbook</strong></a> — get the Tuesday Digest, Thursday Weekly Framework, and free financial tools designed to help you make better decisions about investing, money, and building wealth.</p><p><strong>Learn the framework. Run the model. Make the decision.</strong></p><p>🎥 <strong>Watch </strong><a href="https://www.youtube.com/channel/UCLcEzBcpLsD8cDUZGG246DQ"><strong>Investor’s Handbook on YouTube</strong></a> · 𝕏 <strong>Follow </strong><a href="https://x.com/InvestorsHndbk"><strong>Investor’s Handbook on X</strong></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=205e36a27e2e" width="1" height="1" alt=""><hr><p><a href="https://medium.com/the-alternative-ledger/six-builds-retired-what-the-graveyard-cost-205e36a27e2e">Six Builds, Retired. What the Graveyard Cost.</a> was originally published in <a href="https://medium.com/the-alternative-ledger">The Alternative Ledger</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[You Can Have the Tax Shelter or You Can Have the Keys]]></title>
            <link>https://medium.com/@creed4677/you-can-have-the-tax-shelter-or-you-can-have-the-keys-a3307cdbea55?source=rss-7c07f4a4f0a5------2</link>
            <guid isPermaLink="false">https://medium.com/p/a3307cdbea55</guid>
            <category><![CDATA[financial-planning]]></category>
            <category><![CDATA[personal-finance]]></category>
            <category><![CDATA[investing]]></category>
            <category><![CDATA[retirement]]></category>
            <category><![CDATA[bitcoin]]></category>
            <dc:creator><![CDATA[Colin Reed]]></dc:creator>
            <pubDate>Fri, 30 Oct 2026 15:16:01 GMT</pubDate>
            <atom:updated>2026-10-30T15:44:52.329Z</atom:updated>
            <content:encoded><![CDATA[<p>You Can Have the Tax Shelter or You Can Have the Keys</p><figure><img alt="A worn antique iron key resting on a thick stack of US dollar bills." src="https://cdn-images-1.medium.com/max/1024/0*c7i-2XHpJwmA7PLw" /><figcaption>Photo by <a href="https://unsplash.com/@elibetvm?utm_source=medium&amp;utm_medium=referral">Elibet Valencia Muñoz</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Back in the summer I wrote about whether Bitcoin belongs in an IRA, and I walked through the costs and benefits one line at a time. A few people wrote back with the same follow-up. They weren’t asking about fees. They were asking about the part that doesn’t fit on a fee schedule: if the whole point of Bitcoin is that you can hold it yourself, what exactly are you holding inside an IRA?</p><p>It’s a fair question, and it deserves its own answer. The short version is that the tax shelter and the keys are two different goods, and for the most part you have to pick which one you want each dollar to buy.</p><p>This isn’t investment, tax, or legal advice, and it isn’t a prediction about any price. It’s a way of framing a trade-off so you can make it on purpose.</p><h3>What “the keys” actually means</h3><p>Bitcoin ownership comes down to one thing: who can sign a transaction that moves the coins. Whoever controls the private keys controls the Bitcoin. Not the person named on a statement, not the person with a login, but the person, or the institution, that can produce the signature.</p><p>That’s the property people mean when they say “not your keys, not your coins.” Strip away the slogan and it’s a plain description of counterparty risk. If someone else holds the keys, you hold a claim against them. A claim can be honored, delayed, frozen, or lost, depending on the health and the incentives of whoever owes it to you.</p><p>That doesn’t make a claim worthless. Most of what we own in a retirement plan is a claim of some kind. But it’s a different thing from direct ownership, and Bitcoin is unusual in that direct ownership is actually available to an ordinary person. That availability is a large part of why people hold it.</p><h3>What the shelter actually buys</h3><p>The tax shelter is equally concrete. Inside a traditional IRA, growth compounds without annual tax, and the bill arrives as ordinary income when you withdraw. Inside a Roth, qualified withdrawals come out untaxed. Either way, you can also rebalance inside the account without creating a taxable event each time you trim.</p><p>For an asset someone expects to hold for decades, those are meaningful benefits. I’m not going to argue they aren’t. The question isn’t whether the shelter is valuable. It’s what the shelter costs you in control, and whether that price is worth paying on the dollars you’re considering.</p><h3>Why you can’t have both, mostly</h3><p>Federal rules require IRA assets to be held by a qualified custodian. That’s the structural reason the two goods separate. A custodian holding your IRA’s Bitcoin holds the keys, because holding the assets is what the role is. You can’t point an IRA at a hardware wallet in your desk drawer and keep the account’s tax treatment.</p><p>Some structures exist that try to narrow the gap. Certain self-directed arrangements let an account owner direct an entity that holds the assets, sometimes with more say over how the keys are stored. These are specialized, they carry their own compliance risk and their own costs, and the rules around them are easy to get wrong in ways that can disqualify the account. If you’re looking at one, that’s a conversation for a tax professional who works with self-directed accounts, not a blog post.</p><p>For most people, the practical picture is simpler. Inside the account, a custodian holds the keys. Outside the account, you can. The shelter and the keys live on opposite sides of a line.</p><h3>Three things the custody trade quietly changes</h3><p><strong>Your exposure to someone else’s failure.</strong> Self-custody has its own failure modes: lost keys, bad backups, mistakes made under stress. I’m not pretending otherwise. But custodial risk is a different kind of risk. It depends on a company’s solvency, its controls, its legal structure, and its relationship with regulators. You can diligence those things, but you can’t remove them.</p><p><strong>Your ability to act on your own schedule.</strong> A custodian processes requests on its timeline, within its policies, and inside the account’s distribution rules. Money held in the account is retirement money, with the access restrictions that implies. Money you hold yourself is as available as you decide it is, which is both a benefit and a responsibility.</p><p><strong>Your plan for what happens when you’re gone.</strong> Heirs inherit an IRA through a beneficiary designation and a custodian’s paperwork process. They inherit self-custodied Bitcoin only if they can find and use the keys. Neither is simple, and each fails in its own way. I wrote about the sequencing problem on the self-custody side in a <a href="https://modernwealthmodel.com/bitcoin-inheritance-estate-planning/">separate piece on Bitcoin inheritance</a>, but the point here is narrower: the two structures don’t just differ at the moment you hold them. They differ at every handoff.</p><h3>A decision test you can run in an afternoon</h3><p>Rather than asking “IRA or self-custody?” as a single yes-or-no, split the question into pieces you can answer.</p><p><strong>First: what is this specific dollar for?</strong> Money you expect to leave untouched for twenty or thirty years in a retirement account is the strongest candidate for the shelter. Money you’d want to move on short notice, or money you hold specifically because no third party can freeze it, is the strongest candidate for the keys.</p><p><strong>Second: would you still hold it if a custodian were the only way to hold it?</strong> This is the honest-motive question. If your answer is “yes, I just want the best long-run tax treatment,” the shelter makes sense. If your answer is “no, the self-custody is the reason I own it at all,” then putting that dollar in the account contradicts your own thesis, however good the tax math looks.</p><p><strong>Third: what’s the cost of being wrong in each direction?</strong> A custodial failure and a lost key are both possible. Estimate which one you’re better equipped to prevent. Someone with a tested backup process and a trusted co-signer has a different risk profile than someone who’s never practiced a recovery.</p><p><strong>Fourth: could a split do the job?</strong> Often it can. A common structure is to hold part of the position in the account for the long-horizon allocation, and part in self-custody for control and optionality. The right proportion is personal. But the existence of the split means you don’t have to resolve the question absolutely, and you can revisit the proportion as your situation changes. I keep the mechanics of the account side on the <a href="https://modernwealthmodel.com/bitcoin-ira/">Bitcoin IRA page</a> and the custody side on a <a href="https://modernwealthmodel.com/bitcoin-self-custody-retirement/">self-custody retirement page</a>, so you can compare them without a sales pitch in the way.</p><h3>The part marketing leaves out</h3><p>Nobody selling an IRA wants to lead with “you won’t hold the keys.” And nobody selling a hardware wallet wants to lead with “you’ll give up a tax shelter.” Each side tells the half of the trade that flatters its product.</p><p>The honest version is that both goods are real and both have a price. The shelter is worth the most to long horizons, large expected gains, and active rebalancing. The keys are worth the most to people who hold Bitcoin precisely because it removes a counterparty. Most plans end up using some of each, and the work is deciding how much.</p><p>If you’d rather work through your own version of this with someone than alone against two marketing pages, <a href="https://modernwealthmodel.com/consulting/">Modern Wealth Consulting</a> does fee-for-time Bitcoin retirement advisory. I’m not a custodian, not a tax professional, and not a registered investment advisor. I just spend a lot of time mapping where these structures help, where they quietly cost you, and writing the work down.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a3307cdbea55" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I Sent My Best Side Income to Voicemail for Three Years]]></title>
            <link>https://medium.com/the-alternative-ledger/i-sent-my-best-side-income-to-voicemail-for-three-years-f34d7a265197?source=rss-7c07f4a4f0a5------2</link>
            <guid isPermaLink="false">https://medium.com/p/f34d7a265197</guid>
            <category><![CDATA[entrepreneurship]]></category>
            <category><![CDATA[personal-finance]]></category>
            <category><![CDATA[side-hustle]]></category>
            <category><![CDATA[consulting]]></category>
            <category><![CDATA[careers]]></category>
            <dc:creator><![CDATA[Colin Reed]]></dc:creator>
            <pubDate>Thu, 29 Oct 2026 19:19:22 GMT</pubDate>
            <atom:updated>2026-10-29T19:19:22.823Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="A black rotary telephone with its handset off the hook on a white background" src="https://cdn-images-1.medium.com/max/1024/0*tPwfGQh1ZlCQAZxP" /><figcaption>Photo by <a href="https://unsplash.com/@quinoal?utm_source=medium&amp;utm_medium=referral">Quino Al</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Somewhere around 2019, when I was running operations for an airline caterer, my phone started collecting voicemails from a company I’d never heard of. I read them through the voice-to-text transcription. They asked general questions about airline catering and whether I’d be open to a conversation.</p><p>I assumed it was spam. I never once called back. I never once thought to. A lot of the time I blocked the number so the calls would stop.</p><p>That company was an expert network. It was trying to pay me to talk about the job I was already doing. Years later I’m leaning on platforms just like it for leads, revenue, and a wider view of my own industry than I ever had from inside one building.</p><h3>How I actually found my way in</h3><p>It didn’t start with expertise. It started with a mobile game ad.</p><p>After a layoff in 2025 I went looking for anything that paid. That led to survey sites, and survey sites led to a discovery: if you upload a real resume to a research platform, you stop getting generic surveys and start getting short, specific projects about the work you’ve actually done. Fifteen minutes on a transportation question. Thirty minutes on a supply chain process. Better pay for narrower knowledge.</p><p>Eventually I asked a chatbot a simple question: what else is out there like this? The list it gave back included expert networks, a whole industry I’d technically heard from years earlier and never understood. I applied to the biggest one I could find. Getting picked up was easier than I expected. My first call came in March 2026, on a topic straight out of my last job. The second came two days later, accepted and held the same afternoon.</p><p>Since then I’ve done about 80 calls across 13 networks. That first platform is still my most active.</p><h3>What the work actually is</h3><p>It starts with a screener: a few questions about an industry or a topic. Some are online forms. On some networks a person calls you.</p><p>The person on the other end of a screener is usually an analyst in their early twenties, often a year or two out of school. They’re trained in frameworks, decision analysis, Excel, and research tools. What they don’t have is time on the floor. They’re taking what the client wants and what you say you know and stitching the two together, like medieval monks copying a text they can’t fully read. They don’t fully know what the client is after, and they don’t fully know what your experience involved. They’re doing their best to make a match that holds.</p><p>Once you’re accepted, the other end of the call is the client: an investment firm building a thesis, a private equity firm sizing up an acquisition, a company doing diligence on a merger, a corporate team doing competitive intelligence on a rival.</p><h3>You know more than they do</h3><p>A good call feels easy.</p><p>Early on I over-prepared. I wrote prep documents for every call and tried to anticipate every possible angle. Then I’d get on the call and realize I didn’t need any of it. I could have talked with nothing on my screen.</p><p>That’s the part new experts forget. If the client knew more than you, they wouldn’t be paying a network to find you. Most of the time, I’d say 85 to 90 percent, you’re far better equipped to talk about the topic than the person asking. Some topics do deserve real preparation. The rule still holds.</p><h3>When a call goes wrong, it’s usually the match</h3><p>Bad calls are rarely the client’s fault or the expert’s. Usually the analyst stretched to make a connection that wasn’t there.</p><p>I was once put on a call about getting medicine delivered to homes and care facilities, because my background is delivery and logistics. A few minutes in, it was clear the client expected something I couldn’t give them. So I said so, and offered a choice. We’re already here; let’s make the time count. Let’s talk about cold chain without regard to the product, dry ice, quality standards that carry across industries, liability, fuel surcharge programs, small drops versus large drops, routing, and route density. Or let’s end it, I’ll take the prorated time, and you can find someone who fits.</p><p>Analysts stretch because matching is their job, and some niches are thin. Mine is a strange intersection of airline catering, foodservice, and distribution. Not many people can speak to it, which is exactly why I get pulled into topics most people have never heard of.</p><h3>What gets you picked</h3><p>Keyword density. On your LinkedIn profile and on every network profile you keep.</p><p>Then rotate those keywords with the calendar: seasonality, earnings reports, mergers and acquisitions, headline news, the macro picture, and sometimes geopolitics. Disruption around the Strait of Hormuz pushed plastic resin prices sharply higher this year. Add state-by-state bans on certain disposables, and the takeout container at your local restaurant becomes an investment question. Somebody has to explain that niche to an investor. The network has to be able to find them.</p><p>Your profile is how they find you. Write it so it says plainly what you can speak to and what you can’t.</p><h3>Is this an alternative asset?</h3><p>I’ve come to think of it as one, with limits.</p><p>What an operator knows is a walled garden. The only ways in have been a paid call, a contract, or a full-time hire, and every one is priced by the hour or by the year. That price exists because you can’t search, scrape, or research your way to what someone learned running a fleet or a kitchen. Expert networks are, in effect, the rental desk for that asset.</p><p>The asset is also changing shape. Experts are starting to show up with tools, not just answers. I built a weekly industry digest for myself, and when a client asked how it worked, the network had no box for an expert handing over a system instead of an hour. I’ve also started capturing what I know by talking it into a folder after calls: what I remembered about my own work, never anything a client said. That archive compounds. Every call pulls something out of storage, and every entry makes the next call easier.</p><p>Where it breaks: the income is lumpy and nothing about it is passive. The platforms own the client relationship. Compliance limits what you can say and build. And it’s capped by your hours. It’s an asset you rent out, not one that earns while you sleep.</p><h3>Everybody should try this</h3><p>If you have a domain of expertise at any level of authority, there’s probably a network that wants you. There are a lot of them, with different specialties. I don’t see a reason not to try.</p><p>One test before you start. Go through your blocked and spam calls and look for names you don’t recognize. Mine sat there for three years.</p><p><em>Colin Reed writes about sound money, the real cost of things, and what one operator can build with AI at </em><a href="https://modernwealthmodel.com/"><em>Modern Wealth Model</em></a><em>. More on the expert network side: </em><a href="https://modernwealthmodel.com/expert-network-compliance-ai-tools/"><em>My Expert Network Had No Box for the Tool I Built</em></a><em> and </em><a href="https://modernwealthmodel.com/expert-knowledge-capture-ai-briefing/"><em>The Briefing No One Else Can Write</em></a><em>.</em></p><p>📩 <a href="https://investorshandbook-newsletter-c72f94.beehiiv.com/"><strong>Join Investor’s Handbook</strong></a> — get the Tuesday Digest, Thursday Weekly Framework, and free financial tools designed to help you make better decisions about investing, money, and building wealth.</p><p><strong>Learn the framework. Run the model. Make the decision.</strong></p><p>🎥 <strong>Watch </strong><a href="https://www.youtube.com/channel/UCLcEzBcpLsD8cDUZGG246DQ"><strong>Investor’s Handbook on YouTube</strong></a> · 𝕏 <strong>Follow </strong><a href="https://x.com/InvestorsHndbk"><strong>Investor’s Handbook on X</strong></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f34d7a265197" width="1" height="1" alt=""><hr><p><a href="https://medium.com/the-alternative-ledger/i-sent-my-best-side-income-to-voicemail-for-three-years-f34d7a265197">I Sent My Best Side Income to Voicemail for Three Years</a> was originally published in <a href="https://medium.com/the-alternative-ledger">The Alternative Ledger</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Briefing No One Else Can Write]]></title>
            <link>https://medium.com/@creed4677/the-briefing-no-one-else-can-write-107995fd4565?source=rss-7c07f4a4f0a5------2</link>
            <guid isPermaLink="false">https://medium.com/p/107995fd4565</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[consulting]]></category>
            <category><![CDATA[entrepreneurship]]></category>
            <category><![CDATA[knowledge-management]]></category>
            <category><![CDATA[productivity]]></category>
            <dc:creator><![CDATA[Colin Reed]]></dc:creator>
            <pubDate>Thu, 29 Oct 2026 14:31:46 GMT</pubDate>
            <atom:updated>2026-11-04T19:17:54.533Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="Navy Modern Wealth Model title card reading The Briefing No One Else Can Write, with an orange diamond icon" src="https://cdn-images-1.medium.com/max/1024/0*GVQp-RyOOHjF_zcy.png" /></figure><p>On July 30, 2026, I opened a folder for what I know. Not what I read. What I know: how a fleet actually gets routed before the trucks roll, why a fill rate looks fine on a scorecard and terrible at the back door, what a catering contract really prices. As of October 29, 2026, that folder holds 102 entries across eleven subject areas, and almost none of it exists anywhere else.</p><p>That last part is the point of this piece.</p><p>In <a href="https://modernwealthmodel.com/expert-network-compliance-ai-tools/">Part 1</a>, I wrote about a weekly digest I built that reads public sources and turns them into a five-minute briefing. A client wanted to know how it worked, and the expert network in the middle had no process for an expert handing over a tool instead of an answer. That digest is useful, but anyone with the same source list could rebuild it. This one, nobody can.</p><h3>What I know is a walled garden</h3><p>There have always been exactly a few ways to get at what lives in an experienced operator’s head. You book a paid expert call. You hire the person on a contract. You hire them full time. That’s the list, short of something illegal.</p><p>Every one of those routes is priced by the hour or by the year. Third-party estimates put what clients pay the big expert networks at roughly <a href="https://www.koji.so/blog/expert-networks-vs-customer-interviews-2026">$1,200 to $2,000 an hour</a>. That price exists because the knowledge can’t be searched, scraped, or researched into existence. There are no citations for what I learned running a unionized fleet. No amount of desk research gets you there. The pieces I write here on <a href="https://modernwealthmodel.com/wms-go-live-productivity-dip/">warehouse system go-lives</a> or <a href="https://modernwealthmodel.com/cost-to-serve-new-account/">what a new account really costs to serve</a> draw on exactly that, and you can’t look any of it up. It’s the definition of a walled garden.</p><p>What’s new is a fourth route that doesn’t require me in the room.</p><h3>Talking it out is enough</h3><p>I don’t type most of it. I talk. A debrief on a walk, a voice note in the car, a thought about something I watched go wrong years ago that suddenly matters again. An agent sorts each one into a subject area, tags it, and files it as an entry: what I know, where it applies, how it usually fails, what good looks like, and the questions it equips me to answer.</p><p>The cost is the same AI subscription that runs <a href="https://modernwealthmodel.com/one-man-five-man-team/">everything else I’ve built</a>. <a href="https://modernwealthmodel.com/what-ai-actually-costs-one-person-business/">My software bill</a> didn’t move when the folder started.</p><h3>The work pulls it back out</h3><p>People ask how many hours this has taken. I can’t give an honest number, and the reason turned out to be the interesting part. What’s in the folder isn’t three months of work. It’s eighteen years, coming back a piece at a time.</p><p>Expert calls arrive with little warning about the topic. A question about something I haven’t touched in a decade pulls it out of storage, and a little research reminds me how much I actually did. Most of it I would never have thought to write down on my own. Afterward, I talk it into the folder before it slips away again. What goes in is what I remembered about my own work, never anything a client said or asked.</p><p>So the work feeds the archive, and the archive feeds the work. The longer I do this, the more comes back, and the more that comes back, the more there is to draw on.</p><p>The two systems already talk to each other. Every Monday, the public digest from Part 1 flags a few items worth keeping, and the ones I approve become entries. Public news goes in one side. My judgment about what it means gets saved on the other.</p><h3>A briefing only one person could write</h3><p>Here is the idea that won’t leave me alone. If the digest can turn public sources into a weekly briefing, the same machinery can turn my own entries into one. Not news. Experience, pointed at whatever changed this week.</p><p>A weekly note on what a new routing tool will break, written from what I watched break before. A monthly refresher for a junior operator on how a distributor actually earns money. A one-time packet for an expert network on a topic I know cold, built in an afternoon instead of six calls.</p><p>The uses multiply quickly:</p><ul><li>Expert to expert: two operators trading what they know without booking each other’s time.</li><li>Refresh and recall: keeping a domain I left years ago sharp enough to speak to on short notice.</li><li>A company brain: the same capture habit across a team, so onboarding, consistency, and compliance stop living in one veteran’s head, with tiers that decide which associates can see which parts of it.</li><li>Packaged products: a briefing, a primer, or a license, sold on a cadence instead of by the hour.</li></ul><p>And it compounds. Every entry gets updated as the industry changes, so in ten years the folder isn’t a snapshot of what I knew in 2026. It’s a record of how my thinking moved as the technology did.</p><p>The physicist David Deutsch argues in <em>The Beginning of Infinity</em> that the growth of knowledge has no built-in ceiling, that even the thin hydrogen drifting between galaxies is raw material enough for a civilization that keeps creating it. My folder is a very small version of the same bet. Knowledge isn’t a fixed stock you either have or lose. It’s a process, and a process can be kept running.</p><p>That’s also why I think of this as a question about scarce assets, not just software. Time is the one thing none of us can store. Every year spent running a dock or negotiating a contract is gone once it’s spent. What I learned while spending it doesn’t have to be. Capturing it is the closest thing I’ve found to keeping some of that time.</p><h3>The case against it</h3><p>The strongest objection is old and good. The philosopher Michael Polanyi put it as “we can know more than we can tell.” The most valuable part of experience is judgment in the moment, and a folder of written entries captures the part that could be written down, which is not the same thing. A briefing built from it might read well and still miss what made the operator worth calling.</p><p>There’s a second objection, and it’s mine as much as anyone’s: a briefing with no citations asks the reader to trust one person completely. Part 1’s digest links every claim. This can’t. If I’m wrong, nothing on the page will catch it.</p><p>I think both are true, and I think they set the price rather than kill the idea. A captured brain won’t replace the call. It reaches the people who were never going to book one.</p><h3>Where it breaks so far</h3><p>The folder is three months old, and the cracks are already visible.</p><p>Only 26 of the 102 entries carry the full set of tags that say where each one came from and where it can be used. The other 76 predate that habit and haven’t been brought up to standard. One subject area’s index listed four entries when five were sitting in the folder, and the running count in my own notes drifted before I started recounting from the files themselves.</p><p>The harder problem isn’t bookkeeping. It’s sorting what is mine to sell from what isn’t. Years of experience at former employers, plus work done for clients, all end up in the same head. A briefing for sale can only be built from the part that belongs to me alone, and drawing that line entry by entry is slow, careful work that no agent should do unchecked. Until that sort is done, nothing in the folder leaves it.</p><p>And I don’t know yet whether anyone will pay for a briefing built from one person’s head on a schedule. Part 1 showed that a network wasn’t ready to sell an expert’s tool. This is a stranger product than that.</p><p>Written by Colin Reed</p><p>Colin Reed is an operations executive and independent consultant whose career spans foodservice distribution, airline catering, last-mile delivery, and hospitality, from restaurants and bars to luxury resort food and beverage programs. He has held distribution P&amp;L with capital authority and run enterprise system selections across warehouse management, labor management, and ERP platforms, and advises funds, corporates, and management consultancies through expert networks and research firms. On Modern Wealth Model he writes on applied AI in the physical economy, Bitcoin, and long-horizon retirement planning — and builds the calculators.</p><p><em>Originally published at </em><a href="https://modernwealthmodel.com/expert-knowledge-capture-ai-briefing/"><em>https://modernwealthmodel.com</em></a><em> on October 29, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=107995fd4565" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I Said 20 Hours a Week. My Logs Say 6 to 16.]]></title>
            <link>https://medium.com/the-alternative-ledger/i-said-20-hours-a-week-my-logs-say-6-to-16-ffa5214735d2?source=rss-7c07f4a4f0a5------2</link>
            <guid isPermaLink="false">https://medium.com/p/ffa5214735d2</guid>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[entrepreneurship]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[time-management]]></category>
            <dc:creator><![CDATA[Colin Reed]]></dc:creator>
            <pubDate>Sun, 25 Oct 2026 16:58:47 GMT</pubDate>
            <atom:updated>2026-10-31T16:51:28.466Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="Black flip clock with white numbers against a pale gray wall" src="https://cdn-images-1.medium.com/max/1024/0*t1GCwMUA74O5Iojn" /><figcaption>Photo by <a href="https://unsplash.com/@loic?utm_source=medium&amp;utm_medium=referral">Djim Loic</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Since February I&#39;ve been building a small company&#39;s worth of things by myself with AI agents: a Bitcoin retirement model, a regional sports directory, a site about what food really costs, a small legal tool, and the dashboard that ties them together. Nights and weekends, alongside my consulting work.</p><p>When people ask what that takes, I&#39;ve always given the same answer: about 20 hours a week. Now I&#39;m writing that answer into a case study. Twenty hours a week since February comes to roughly 700 hours, and I want to set that next to what the same work would have taken a team without AI. That comparison is only as honest as the 700 underneath it.</p><p>So on October 6 I wrote the number down as a baseline. Two weeks later I checked it against the record. The record disagreed.</p><h3>I should have known better</h3><p>I came up in food distribution, where nobody&#39;s time is taken on their word. A labor management system puts an engineered standard on the work, cases picked, pallets loaded, stops run, and a selector&#39;s shift is measured against that standard, not against how the shift felt. A good supervisor could tell you within a few minutes what a route should take. A weak one told you how hard everybody was working.</p><p>Then I started building on my own time and did exactly what I used to correct in other people. I estimated. I never measured.</p><h3>How I rebuilt two weeks</h3><p>There&#39;s no time clock on a kitchen-table operation, but there is a paper trail. Every project keeps a dated build log, and the entries show when I was actually in the room: an approval, an instruction in my own words, a decision only I could make. Jobs that ran on a schedule while I slept count as zero.</p><p>Then a rough scale, the same way I used to rough out a shift before the reports came in. A light day with one item is half an hour to an hour. Two to four items is one to two hours. Five or more is two to four.</p><p>The week of October 5 had four working days: a new series linked across the site, two social accounts going live, fixes to the legal tool, and a published essay. It came to 6 to 10 hours.</p><p>The week of October 12 had six working days, including the heaviest of the month, when I published an episode, ran a style pass across a whole site, and put out two more posts. It came to 10 to 16 hours.</p><p>I read both tables line by line and signed off on them. Neither week reaches 20.</p><h3>The case for my original number</h3><p>Here&#39;s the best defense of 20 hours. The logs only see time that changed a file. They don&#39;t see me reading a draft on my phone at nine at night, the half hour in the car deciding whether a series should exist at all, or the session where I went back and forth with an agent for an hour and threw the whole thing away. In a warehouse that&#39;s indirect time, and every labor standard carries an allowance for it. My reconstruction carries none.</p><p>So 6 to 16 is a floor, not a count. The real number sits somewhere above it, and I can&#39;t tell you how far.</p><h3>Why the gap matters</h3><p>The convenient part is that fewer hours make the story better. One operator, no development background, a small company&#39;s worth of software, content, and operations, for less time than I claimed. I don&#39;t get to pick the version I like, though. The reason to keep a ledger is that it&#39;s allowed to embarrass me.</p><p>The 700 hours I started with is just 20 a week multiplied out. A separate per-project estimate built from file dates lands in the same place, and I&#39;d love to call that confirmation. It isn&#39;t. Both lean on the same assumption about how I work, and two estimates agreeing doesn&#39;t turn either one into a measurement.</p><h3>Where this leaves me</h3><p>Two weeks is a small sample, and a reconstruction is still an estimate with better manners. What I&#39;m sure of is narrower: the number I&#39;d been quoting for eight months had never been checked once. On a distribution floor, that would have been the first thing I flagged.</p><p>So the 700 hours doesn&#39;t go into the case study as a fact. It goes in as what it is, an estimate, next to a log that started October 6 and is still running. When someone asks me what this takes, the honest answer for now is less than I said, more than I can prove, and I&#39;m counting.</p><p>Colin Reed writes about Bitcoin retirement, the real cost of things, and what one operator can build with AI at <a href="https://modernwealthmodel.com/">Modern Wealth Model</a>.</p><p>📩 <a href="https://investorshandbook-newsletter-c72f94.beehiiv.com/"><strong>Join Investor’s Handbook</strong></a> — get the Tuesday Digest, Thursday Weekly Framework, and free financial tools designed to help you make better decisions about investing, money, and building wealth.</p><p><strong>Learn the framework. Run the model. Make the decision.</strong></p><p>🎥 <strong>Watch </strong><a href="https://www.youtube.com/channel/UCLcEzBcpLsD8cDUZGG246DQ"><strong>Investor’s Handbook on YouTube</strong></a> · 𝕏 <strong>Follow </strong><a href="https://x.com/InvestorsHndbk"><strong>Investor’s Handbook on X</strong></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ffa5214735d2" width="1" height="1" alt=""><hr><p><a href="https://medium.com/the-alternative-ledger/i-said-20-hours-a-week-my-logs-say-6-to-16-ffa5214735d2">I Said 20 Hours a Week. My Logs Say 6 to 16.</a> was originally published in <a href="https://medium.com/the-alternative-ledger">The Alternative Ledger</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[My Expert Network Had No Box for the Tool I Built]]></title>
            <link>https://medium.com/@creed4677/my-expert-network-had-no-box-for-the-tool-i-built-2efa78bc52cc?source=rss-7c07f4a4f0a5------2</link>
            <guid isPermaLink="false">https://medium.com/p/2efa78bc52cc</guid>
            <category><![CDATA[consulting]]></category>
            <category><![CDATA[entrepreneurship]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[future-of-work]]></category>
            <category><![CDATA[compliance]]></category>
            <dc:creator><![CDATA[Colin Reed]]></dc:creator>
            <pubDate>Thu, 22 Oct 2026 14:17:24 GMT</pubDate>
            <atom:updated>2026-10-27T18:19:16.903Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*zBIsgBlr6Ggbqd9_.png" /></figure><p>In October 2026, near the end of an ordinary paid call through an expert network, a client asked how I stay current on the industry. My answer turned into the first request I’ve seen that the network’s compliance team had no process for.</p><p>I didn’t answer with a list of trade publications. I told the client about a weekly digest I built for myself. It’s an AI agent that reads a curated set of sources early every Monday, sorts the prior week into fixed lanes, writes a five-minute briefing with a source link on every item, and drops it in my inbox. It runs on my own setup and on the AI subscription I was already paying for. It added nothing to <a href="https://modernwealthmodel.com/what-ai-actually-costs-one-person-business/">my monthly software bill</a>.</p><p>The client wanted to see how it worked. The network set up a short paid written engagement, and I delivered a plain overview: what the digest does, the steps it runs, the design choices that keep it short, and the kinds of sources it reads. I also listed what a deeper build would involve.</p><p>Then the follow-up stalled.</p><p>For several business days, the question of how to take it further sat with the network’s compliance team. The account rep apologized more than once and was straight with me: they had never handled a request like this, and they were building the process as they went.</p><p>Even the part that was already finished got tangled. Over the following week, I was asked for the same written work four separate times, once in the original request and three more times through automated emails, and I submitted it twice. The work hadn’t changed. The system just didn’t seem to know where it was supposed to go.</p><h3>Nothing about the request was hard</h3><p>From my side, this was easy work. The system already exists. Pointing it at a different industry is mostly a matter of changing the source list and the lanes.</p><p>From the client’s side, it was a small ask. They wanted to know how one operator keeps up, and whether their team could do the same.</p><p>And the network wanted the work. Nobody in this story was trying to slow it down. Everyone wanted the same outcome, and the process had no slot for it.</p><h3>The network sells what I know, not what I built</h3><p>Expert networks are built around a clean idea. A client pays for an hour of someone’s knowledge. The expert talks, the call ends, and the transfer is over. Compliance is built to police that exchange: nothing confidential from a current or former employer, nothing material and non-public, no side deals.</p><p>A written answer fits that model. So does a survey.</p><p>A working system doesn’t. Once the client asked how they could run something like my digest themselves, the request stopped being about what I know and became about something I made. That raises questions an hourly-call process was never designed to answer. Who owns a build guide? Is it a consultation or a license? How do you price an hour when the deliverable is a method someone keeps using for years? Does handing it over start an ongoing relationship the network now has to watch?</p><p>I don’t know how they’ll answer those. I suspect they don’t yet either. That is the whole point.</p><h3>The case for going slow</h3><p>The strongest argument against my impatience is that compliance is doing exactly its job. The expert network industry went through insider-trading prosecutions a little over a decade ago, and the business rests on clients trusting that nothing improper passes through a call. A new kind of deliverable is precisely where a careful compliance team should slow down. If I ran that team, I would want to see the thing before I priced it, too.</p><p>I agree with all of that. My point is narrower. The delay isn’t a sign that anyone did something wrong. It’s a sign that a category is missing.</p><h3>Experts are starting to show up with tools</h3><p>Here is what I think is changing. Until recently, an operator showed up to an expert call with experience and a point of view. Now some of us show up with experience, a point of view, and software we built for our own practice.</p><p>I have no developer background. Since February 2026 I’ve built, with AI agents, a <a href="https://modernwealthmodel.com/one-man-five-man-team/">small company’s worth of tools</a>, mostly for my own use. The digest is one of the smallest. It does a job I would otherwise cover with a paid monitoring service or with hours of my own reading every week.</p><p>I won’t be unusual for long. When a retired plant manager, a former buyer, or a one-time finance lead can build a research tool in an evening for close to nothing, the question “how do you stay current?” starts to have answers that are things, not habits. Clients will ask for the things.</p><p>Networks will need a box for that: a way to scope it, price it, check it for anything that should never leave the expert’s head, and hand it over cleanly. I expect the first network that builds that box well will win work the others end up turning away.</p><h3>Where I’m drawing my own lines</h3><p>Until that box exists, I keep a few rules for myself.</p><ul><li>I describe the system, never the client work it came out of. This post names no network, no client, and nothing said on the call.</li><li>The full source watchlist and the working instructions stay mine. What I hand over explains how it works and why, at a level a team can act on.</li><li>Nothing in the tool, or in anything I deliver about it, comes from a former employer. It reads public publications and company pages.</li><li>Any version I build for a client I met through a network goes through that network.</li></ul><h3>Where this breaks</h3><p>This is one request, at one network, in one industry. One slow ticket is an anecdote, not a trend, and I may be reading too much into a slow week.</p><p>It did end. A little over a week after I first sent the written work, I followed up once more, someone more senior stepped in, and the payment went through shortly after. What I still don’t know is whether the follow-on, the part that would have meant handing over the system itself, will ever get a process of its own.</p><p>It also leans on an easy case. A news digest reads public sources and nothing else. A tool built on a client’s own data, or on anything close to an expert’s former employer, is a much harder handoff, and a network would be right to move even slower with it.</p><p>There’s a bigger version of this that I keep coming back to. My digest reads public sources. The same machinery could read an expert’s own captured knowledge instead: twenty years of decisions, failures, and fixes, saved as they happen and turned into a recurring briefing. That’s a very different tool, and a much harder compliance question. It’s where I’m headed next.</p><p>If you’re an operator thinking about building something like this for your own practice, <a href="https://modernwealthmodel.com/ai-consulting/">the intro call is free and fifteen minutes</a>.</p><p>Written by Colin Reed</p><p>Colin Reed is an operations executive and independent consultant whose career spans foodservice distribution, airline catering, last-mile delivery, and hospitality, from restaurants and bars to luxury resort food and beverage programs. He has held distribution P&amp;L with capital authority and run enterprise system selections across warehouse management, labor management, and ERP platforms, and advises funds, corporates, and management consultancies through expert networks and research firms. On Modern Wealth Model he writes on applied AI in the physical economy, Bitcoin, and long-horizon retirement planning — and builds the calculators.</p><p><em>Originally published at </em><a href="https://modernwealthmodel.com/expert-network-compliance-ai-tools/"><em>https://modernwealthmodel.com</em></a><em> on October 22, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2efa78bc52cc" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>